Tuesday, February 22, 2022

 ... by their fruits ye shall know them.

Matthew 7.20

Tom Elliott and Valeria Vitale, two members of the Pleiades team, have left us imaginary visions of what it is that they hope to achieve through their software initiative.  These statements, the first from 2009 and the second from 2021, reveal in  exciting detail what kinds of software they hope to produce.  I  quote a bit from each but their entire statements are worth reading and I urge readers of this post to do so.
  1. "As I settle into my chair, a second cup of morning coffee in my hand, an expansive view of the eastern Mediterranean fades in to cover the blank wall in front of me. It's one of my favorite perspectives: from a viewpoint a thousand kilometers above the Red Sea, you can look north and west across an expanse that encompasses Jordan, Egypt and Libya in the foreground and Tunisia, Calabria and the Crimea along the distorting arc of the horizon. A simple voice command clears some of my default overlays: current precipitation and cloud cover, overnight news hotspots, and a handful of icons that represent colleagues whose profiles indicate they're currently at work. Now I see a new overlay of colored symbols associated with my current work: various research projects, two articles I'm peer-reviewing and various other bits of analysis, coding, writing and reading. These fade slowly to gray, but for two. Both of them are sprawling, irregular splatters and clumps of dots, lines and polygonal shapes." [1]
  1. "As we settle into our chairs for the monthly meeting of the Early Islamic Interest Group, some with a second cup of tea in hand, the Chair’s voice assistant brings up the ‘Pelagios Network spatial feed’. An expansive view of the Mediterranean appears on our screens, with current weather patterns, the latest research publications and colleagues’ avatars briefly appearing above their respective locations – Egypt, Tunisia, Italy and Cyprus among them. These are, however, quickly replaced by a layer of content reflecting the main business of the day: a regional museum in Libya has digitally published their numismatic collections with associated spatial data. Reem Fathi, the museum’s curator, introduces her team’s work and begins by displaying the distribution of the early Islamic coins. Their alignment with settlements along the Cyrenaican coast, hinterland and adjoining islands is immediately apparent. Now the fun begins."  [2]

It is a remarkable fact that when Elliott (1) and Vitale (2) are deprived of the use of computer buzzwords and jargon they are capable of describing what they want with startling clarity.   Two observations:

a. It really would be tremendous to have such powerful tools.

b. How extraordinarily sad it is that they have staked their entire effort on the one software approach that has not produced, and can never produce, the results they want.

Their total and absurd commitment to the concept of the Semantic Net (the computer equivalent of spinning straw into gold) has driven them into an epistemological corner such that after having spent millions of dollars in grant money over the last decade (It was north of 2 million USD when I added it up 4 years ago) they have given the Humanities community three things:

a. bad data (Pleiades)
b. non-working product (Peripleo)
c. do-nothing software (Recogito)

They are now at an impasse.  They have no success to show and no way forward.  Their only hope is to get other projects to re-engineer themselves into Pleiades' semantic net.   By convincing others to change themselves, instead of Pleiades changing itself,  Pleiades gives the illusion of activity and obscures the fact that, as currently managed, it has no hope of accomplishing its goals.

In the meantime the Pleiades team continues to spawn a tornado of papers couched in a nearly impenetrable computer jargon (I give a lexicon of this rubbish here) which gives the illusion to naive Humanists[3] that the Pleiades team is Doing Something Really Important.   Vitale et al. [2021]'s restatement (twelve years later) is nearly identical to Elliott's original thus exposing that their progress towards accomplishing their stated goal has been precisely zero.

How have things come to such a pass?

As a former computer professional myself it's easy to recognize the signs of a disaster in the making.  Aside from the two statements I reproduced above I have seen no other example of their ever having performed what professionals call 'Use-Case Analysis'.  It is here that designers put down in great detail exactly how the final product is to be used and what specific abilities it will give the user/operator.   From the Use Case flows the requirements document which details the specific things that the resulting software must do.  And not only the software is involved but requirements also directly control database design.  Database design is carefully specified in documents of its own and these are keyed to the software documents.  You won't be able to fulfill a DB query if the DB itself doesn't support what's being asked.  The P. team might respond with such piffle as 'that is the old stodgy way of doing things and that such formal documents only get in the way of creativity'.  

But it is exactly the 'compute first, figure it out later' mentality of Pleiades that is the old stodgy way of doing things.  

Understanding the design specifics by creating detailed and shareable documents represents hard-won knowledge consequent to many expensive failures.

The primary problem here is either poor software management or, more likely, none at all.  Lack of transparency, cost overruns, extraordinary amounts of development time, over-promising, under-delivering, etc., etc.  Believe me, I've seen all these pathologies before.

Pleiades' current problem is that the more projects they convince to come on board the more likely it is that these project leaders will start to ask embarrassing questions.   Questions such as 'why should we expose our data for your benefit  and how does this benefit us?', 'Why doesn't Pleiades/Peripleo give useful results even though we expended the time and resources to reverse-engineer our own project to fit it?'  As more time goes by the  leaders of the Pleiades/Peripleo effort will be less and less able to give satisfactory answers to these legitimate questions.

I. The disappointment of Pleiades becomes all the more acute when we compare it to an initiative that deals with exactly the same subject matter, namely, The Levantine Ceramics Project[4], which, as websites go, is a thing of beauty and has not cluttered up the design process with Semantic Net jargon and simply gotten on with it.

II. If the  Pleiades team had chosen a professional approach they might very well have accomplished what they wanted to.

What then must we do?

Questions will inevitably arise concerning what Pleiades is delivering to the Humanities community for all the money that they have spent.  The day of reckoning must eventually come.  There are real reasons for concern that Pleiades is freezing into a kind of fetal crouch.  They have no way forward.  I have heard, although I do not know for a fact, that they have exhausted their funds - which seems incredible to me.  Peripleo, their flagship product, has been down at least since November.  It is down as I write. [Although down through most of February, as of 2/22/2022 at 10:18  PST it has come back up.]  It came up briefly at the end of November and I noticed that all their links to the British Museum database were broken.  This was a good indicator that the system is unstable and breakage-prone. That seems to be what led to their next down-period.  Pleiades appears to have no infrastructure in place in the form of any sort of team that can react to problems in real-time.  We must anticipate further such disruptions in Peripleo's service.  I see on GoFundMe many archaeological projects begging for tiny amounts of money in order to carry on their work; amounts like five hundred dollars to fund the rental of a storage shed for finds.  And yet Pleiades/Peripleo has received millions in grant money to produce products that are inappropriate for scholarly work.

Remedial steps need to be taken and the sooner they are taken the better the possibility that Pleiades can recover.  If nothing is done to restrain them in their course their inevitable embarrassment and collapse will be all the more stark.

I, like Elliott and Vitale, have a vision for the future.  It consists of the following four recommendations.

(a) The Pleiades team needs to explain to the Humanities community exactly what it is trying to do that will benefit the community; e.g. what computer/software faculties are gained for the community which did not previously exist.  It needs to do this in a transparent manner and without resort to the computer buzzwords or formalisms with which it has cloaked its work so far.[5]  The PT needs to prepare a set of Use Cases that detail how their end product will be used. 

(b) The Pleiades team needs to be very clear about the state of the licenses of its  own products as well as the licenses of those who have contributed their data to the Pleiades backbone. 

(c) The  Pleiades team needs to make a public pledge that it will never seek to sell or license the Pleiades technology to any customer, particularly schools, museums, or universities.  In particular it will not bundle and sell data and projects that were donated to them by other teams.  If the ultimate plan of Pleiades is to sell or lease its technology (and, yes, these things have happened) then they need to be transparent about that fact and transparent about what use they intend to make of the data from those forty or so teams that freely contributed to Pleiades/Pelagios/Peripleo.  For example, will their contributors be compensated?

(d) The Humanities community (AIA?, SCS?) needs to appoint an independent working group to evaluate Pleiades' products for utility and likely usefulness for workers and projects in the disciplines of Anthropology, History, Toponymy, Classics, and allied disciplines.  One emphasis should be on how much the product has cost vs. what it has delivered.  When that working group has completed its study it should prepare an open and clear report about what they have learned and make this a formal recommendation to the National Endowment for the Humanities which has provided the bulk of Pleiades' funding so far.

Next time:  What use should be made of Pleiades (non) database?

  1. Elliott and Gillies [2009], first paragraph.
  2. Vitale et al. [2021] 5.  This is the first part of the first paragraph.  
  3. I mean no insult by the phrase 'naive Humanists'.  Indeed, why should workers in the Humanities be expected to be computer experts?    But the inescapable fact that they are not experts makes Humanists an inviting target for schemes such as Pleiades'.
  4. Online here.
  5. See this post for a lexicon of 'Pleiadese'.


Elliott and Gillies [2009] : Elliott, Tom and Sean Gillies.  'Digital Geography and Classics', Digital Humanities Quarterly(3:1) 2009.  Online here.

Vitale et al. [2021] :  Vitale, Valeria and Pau de Soto, Rainer Simon, Elton Barker, Leif Isaksen, Rebecca Kahn.  'PELAGIOS – Connecting Histories of Place. Part I: Methods and Tools'.  International Journal of Humanities and Arts Computing 15.1-2 (2021): 5–32.  Online here.

Sunday, February 20, 2022

An Electronic Barrington's Atlas. A New Feature in the Mycenaean Atlas Project

 The Mycenaean Atlas Project is happy to announce that it is now possible to see several well-known geographic databases side by side on the same map.  These databases, right now, are the complete Pleiades dataset, the harbor dataset from Arthur de Graauw, and the Topostext dataset from Brady Kiesling.  

Here's what the area of the Piraeus looks like on the new page which I call 'Coverage':

Area of the Piraeus.  Click on map to enlarge.

The key here is 'P' : Pleiades, 'H' : Arthur de Graauw's Harbor site, 'T' : Topostext and the blue paddle is from the Mycenaean Atlas Project.  When you mouse over these icons the tool-tip giving the id no. as well as the name string (if there is one) appears.  When you click on an icon you will see an info box with a link to an appropriate page that gives more info about the site.

The Topostext icons provide a link to the relevant Topostext page.  The Pleiades markers link to the appropriate Pleiades page and the M.A.P. markers link to the right page in the Mycenaean Atlas.  The geo-locations marked with an 'H' link to Arthur de Graauw's splash page only.  It is very unfortunate that M. de Graauw does not have a more significant web presence.  In my view he should be able to provide individual pages for each of his harbor sites.  If these existed then I would happily link to them.

Currently your access to this product is through the M.A.P. Place Key Report page.  On that page you pull down the toolbox drop-down and select 'Coverage'.  When you do that the Coverage page appears with your site in the center.  

Arrow pointing to Coverage Page selector

In this picture the arrow points to the 'coverage' selection on the tool box menu.  This is the Place Key Report page.  The lat/lon center of the coverage page will be the lat/lon position of the site currently being examined.  This is site C1232 for Cape Manika on Euboea.

There is a third parameter to the page which governs how wide an area is searched and shown.  This value ranges from 0.01 to 0.09 and represents fractions of degrees.  0.01 searches a rectangle about 2 km in width and 0.09 searches a rectangle about 18 km in width.  The coverage page is currently defaulted to 0.09.  

The ability to display these datasets on the same map creates what all of us have wanted for a long time - a complete Digital Atlas of the ancient world which covers time periods from the Early Bronze Age to the early Medieval period.   And it is fascinating to look at the several perspectives about what such an atlas should include as seen by the different contributors.

Area of Chalcis on Euboea

de Graauw's harbor markes cluster on the coasts or in the watter.  Topoguide seems to have conentrated on Roman ruins in Chalcis.  Pleiades' coverage here is sparse.  Mycenaean Atlas blue paddles are mostly not where anyone else's are (Bronze Age, don't you know?).

Now I need to say  some stuff for legal purposes:

Locations associated with an  are the work of Arthur de Graauw and are used with permission (bottom of page).   His site is called "Ancient Coastal Settlements, Ports and Harbours" and is found here.  When a link associated with a  is clicked on the user will be on the splash page of M. de Graauw's site.

Locations denoted by a  are the work of Brady Kiesling and are used with permission. As he says: "Linked Open Data Files for use with appropriate attribution to ToposText."   His site is called "Topostext" and is found here.  When a link associated with a  is clicked on the user will be on the Topostext site.

Locations denoted by a  are the work of the Pleiades Project and are used with permission. Their site is called "Pleiades" and is found here.  As they say: "Using, sharing, and remixing of the content is permitted under terms of the Creative Commons Attribution 3.0 License (cc-by)."  When a link associated with a  is clicked on the user will be on a Pleiades page.

Each geo-location marker provides a pop-up info box with a link.